This paper presents a few comprehensive experimental studies for automated Structural Damage Detection (SDD) in extreme events using deep learning methods for processing 2D images. In the first study, a 152-layer Residual network (ResNet) is utilized to classify multiple classes in eight SDD tasks, which include identification of scene levels, damage levels, material types, etc. The proposed ResNet achieved high accuracy for each task while the positions of the damage are not identifiable. In the second study, the existing ResNet and a segmentation network (U-Net) are combined into a new pipeline, cascaded networks, for categorizing and locating structural damage. The results show that the accuracy of damage detection is significantly improved compared to only using a segmentation network. In the third and fourth studies, end-to-end networks are developed and tested as a new solution to directly detect cracks and spalling in the image collections of recent large earthquakes. One of the proposed networks can achieve an accuracy above 67.6% for all tested images at various scales and resolutions, and shows its robustness for these human-free detection tasks. As a preliminary field study, we applied the proposed method to detect damage in a concrete structure that was tested to study its progressive collapse performance. The experiments indicate that these solutions for automatic detection of structural damage using deep learning methods are feasible and promising. The training datasets and codes will be made available for the public upon the publication of this paper.
A large number of reinforced concrete (RC) buildings and bridges is deemed structurally deficient. This is either because the infrastructure continues to age and deteriorate or the strength or deformation capacity of the existing older infrastructure does not meet the current code requirements, e.g., in high seismic regions. Thus, the need for more efficient retrofit methods has increased in recent years. Currently, there are only a few methods used for strengthening or retrofitting columns. Steel jackets and Fiber Reinforced Polymer (FRJP) composites are the two most commonly used methods. In this study, along with these two retrofit methods, concrete jackets reinforced with spiral rebar, Welded Wire Fabric (WWF), and a new steel reinforcement termed PCS are investigated under different axial load conditions.
A new concrete confinement model is developed to predict the axial load versus displacement behavior of circular columns under concentric axial load. The new confinement model is proposed for concrete filled steel tube columns as well as circular reinforced concrete columns with steel tube jacketing. Existing confinement models were evaluated and improved using available experimental data from different sets of columns tested under similar loading conditions. The proposed model is based on commonly used confinement models with an emphasis on modifying the effective confining pressure coefficient utilizing the strength of the unconfined concrete and the steel tube, the length of the column, and the thickness of the steel tube. The proposed model predicts the ultimate axial strength and the corresponding strain with an acceptable degree of accuracy while also highlighting the importance of the manner in which the steel tube is used.
There are a large number of reinforced concrete buildings in seismically active areas of the world that are not built in accordance with modern seismic design provisions such as those published by American Concrete Institution ([ACI] Committee 318, 2008). In the United States and other parts of the developed world, these buildings were constructed between 1930s to mid 1970s according to the building code requirements of that time. Even today, in low to moderate seismic regions and in some developing countries that are in process of developing and implementing their seismic codes, reinforced concrete structures are being designed and built without essential seismic details deemed vital to withstand large lateral loads. These buildings often have low lateral displacement capacities and undergo rapid degradation of shear strength and axial load carrying capacity during strong ground motions and hence are extremely vulnerable to excessive structural damage or collapse during future earthquakes. In the past, the earthquakes have caused wide spread damage to the reinforced concrete structures with inadequate seismic design and construction practices. For example, during Kashmir (Pakistan) earthquake of 2005 and Haiti earthquake of 2010, extensive structural damage to residential, commercial and government buildings was observed (Earthquake Engineering Research Institute [EERI], 2005; Mid-America Earthquake [MAE] Center, 2005; U.S. Geological Survey [USGS]/EERI, 2010). The damage was attributed largely to lack of earthquake-resistant design, poor standard of construction and inferior quality of building materials. In majority of the collapsed or damaged structures, structural types, member dimensions and detailing practices (insufficient lap length, improper lap location and lack of confinement in columns etc) were found inadequate to resist forces imposed by these earthquakes. The 2011 off The Pacific Coast of Tohoku (Japan) earthquake is a modern day example of large scale devastation to a highly industrial nation in which building and infrastructure is well designed and constructed. Although, majority of causalities and large scale destruction of infrastructure was caused by ensuing tsunami, limited damage to the buildings due to ground shaking was reported (Pacific Earthquake Engineering Research Center [PEER]/EERI/Geotechnical Extreme-Event Reconnaissance [GEER]/Tsunami Field Investigation Team, 2011; Takewaki et al. 2011). However, extensive and severe structural damage was observed in older residential and commercial buildings that were constructed prior to 1978 code revision of Japan, whereas modern structures built to withstand seismic demands did not sustain any substantial and widespread damage (Aydan & Tano, 2011).
Retrofit and rebuilding are two potential solutions for reinforced concrete (RC) structures with columns that have insufficient detailing or strength to withstand seismic loads or other extreme events. Demolition and rebuilding can be expensive and time-consuming. Thus, retrofitting and strengthening of columns can be an efficient solution. Jacketing of reinforced concrete columns is one of the common methods for retrofit and rebuilding. For jacketing method, the main concern is the performance loss between new and old concrete due to interface slip. There are three major options to provide a better performance, as surface roughening, dowels, and both. In this study, the methods to provide coworking of new and old concrete are evaluated with respect to four parameters; strength, energy dissipation, stiffness, and displacement ductility. Additionally, a slip coefficient is proposed to model the friction between new and old concrete. The proposed model is applied into OpenSees and a Finite Element (FE) software. The slip coefficient is simulated numerically by a friction coefficient using a tangential contact behavior definition as well. Finally, a slip coefficient is proposed for each method.
Collapse performance of nine existing multi-story buildings have been investigated through field experiments and computational modelling. Masonry, reinforced concrete and steel frame buildings have been tested. Most of the test buildings were located on the Ohio State University campus. Single or multiple first-story columns were physically removed from each building during the experiments. The data produced in this research has been a valuable addition to the state of knowledge on progressive collapse of buildings because experimental evidence from full-scale structures is either very limited or does not exist. The main goal of field experiments was to simulate the dynamic and static response of buildings that may experience progressive collapse after sudden loss of column(s). Another objective was to investigate how the internal forces would be redistributed within the building after column loss. Experimental data is used to evaluate the design methods and analysis procedures recommended in current design guidelines. Two and three-dimensional building models were analyzed to simulate the progressive collapse response. Computational models and simulations were compared with the experimental data from the field tests. This study showed robustness of existing buildings and potential contribution of structural components to collapse resistance.
This paper presents the results of field performance tests of 39 in-service corrugated steel highway culverts in Ohio. The culverts had span lengths varying from 3.23m(10.6ft)to7.04m(23.1ft) and backfill soil heights over the crown varying from 0.27m(0.9ft)to7.47m(24.5ft). Static and dynamic load tests were conducted by driving heavy trucks across the culverts. Static loads were applied at ten different locations above each culvert. Dynamic load tests were conducted at six truck speeds varying from 8km∕h(5mi∕h)to64km∕h(40mi∕h). A portable instrumentation frame was installed inside each test culvert to monitor deflections. Strains on the culvert walls were also measured at 14 locations using strain gauges. Effects of backfill height and loading conditions are investigated. According to the experimental results, a plot of maximum culvert deflection versus backfill height shows a nonlinear relationship. Maximum static load deflections were found to be consistently larger than the maximum dynamic deflections obtained using the same test truck. Deflections were nearly zero for deep culverts with backfill heights exceeding 4m(13ft). Maximum deflections correlate more closely to equivalent line loads than to total truck weight. The data also indicate that culvert behavior is more difficult to predict when backfill heights are shallow because other factors, such as culvert age and condition and soil type, likely play a significant role.
A large number of corrugated metal pipe-arch culverts are located under highways. This study investigates the field performance of four existing pipe-arch culverts under static and dynamic loads. Effects of various parameters were considered in selection of the culverts, including backfill height, loading conditions, age of placement, and culvert geometry. Static loads were applied at ten different locations above each culvert using heavily loaded test trucks. Six dynamic tests were conducted at speeds varying from 8 to 64km/h . A portable instrumentation frame was installed inside each test culvert to monitor the deflections at five critical locations. During each test, strains were also measured using 14 strain gauges. Test results indicated that culvert response was influenced significantly by the backfill height. Nearly symmetrical deflection patterns were recorded for symmetrical loading about the longitudinal vertical plane through the crown. The maximum static deflections were larger than the maximum dynamic deflections for each culvert.
Predicting the shear strength of structural elements subjected to gravity loads and ground motions is an important component of seismic design. Among all primary structural components, the vital role of columns in load transfer and redistribution, structural stability, and collapse prevention has been well recognized through observations made in the aftermath of past earthquakes. Numerous analytical, numerical, and experimental studies have been conducted to assess the shear strength of RC columns in the past decades. However, there is still a large scatter (i.e., uncertainty) in the predictions of current empirical and numerical models relative to test data. In this paper, novel data-driven models are presented for predicting the maximum shear strength of rectangular and circular RC columns. To this end, two extensive experimental databases for both types of columns were used in the present study. The data were randomly partitioned into calibration and validation sets. The calibration data sets served as the basis for developing linear and nonlinear equations for predicting the ultimate shear capacity of rectangular and circular RC columns through regression analyses. The Monte Carlo method was employed by conducting 106, 107, and 108 realizations to exhaustively examine the optimal parameter space of the postulated equations. The calibrated predictive models were then validated using the validation data sets; and their performances were also compared to existing models. These validation and comparison studies revealed that a new linear model devised and calibrated in the present study achieved very high accuracy, even compared to various nonlinear models considered. It was, moreover, significantly superior to all prior models in predicting the column shear strengths. This linear model, which is based on physical parameters, can therefore be recommended for engineering practice.
This article, drawn from the National Cooperative Highway Research Program (NCHRP) Synthesis 581: Rehabilitation of Culverts and Buried Storm Drain Pipes, presents current and emerging culvert repair and rehabilitation methodologies, as well as factors affecting their selection and implementation. The synthesis collected and analyzed information from a variety of sources including: a literature review of relevant standards, best practices, and guidance documents; a web-based questionnaire that was completed by 42 state departments of transportation (DOTs); and follow-up interviews to develop case examples. Four major culvert types were examined: corrugated metal, precast concrete three-sided or box structures, precast concrete pipe or arch, and thermoplastic pipe. The survey questionnaire and case examples focused on: culvert types and frequency of their rehabilitation, critical factors affecting the rehabilitation method, local and spot repairs, repair of joints between pipe pieces or metal plates, invert paving and invert lining rehabilitation methods, system-level rehabilitation methods for the entire culvert lining, existing structural capacity, in-house rehabilitation, and quality assurance. This article discusses key findings, spot and joint repair methods, system-level methods, consideration of existing structural capacity, and case examples from various state DOTs.
Parameters affecting the shear strength of reinforced concrete columns having a rectangular cross section and light transverse reinforcement are investigated using data from numerous column tests. A new model is proposed to predict the column shear strength based on theoretical formulations and experimental evidence. The proposed shear strength equation includes contributions from the concrete and transverse reinforcement. Primary parameters in the shear strength model are the column cross-sectional dimensions, concrete compressive strength, column aspect ratio, axial load, and displacement ductility demand. The proposed model is compared with other shear strength models using the available column test data and is shown to result in improved accuracy. A conservative shear strength model for use in design and assessment is proposed based on statistical evaluation of computed and actual shear strengths.
Reinforced concrete jacketing is one of the most frequently used methods for strengthening of reinforced concrete (RC) columns. A large number of experimental studies have been carried out to investigate the effectiveness of repair and strengthening techniques and interface treatment on the response of concrete jacketed columns. However, the effects of potential damage in existing column and quality of core concrete on the response of jacketed RC columns have not been investigated. One of the main goals of this study is to examine how the material properties of the existing column affect the overall response of the jacketed RC columns. Two computer models were developed and nonlinear analyses were performed to determine the moment-curvature relationships and axial load-moment interaction diagrams of concrete jacketed RC cross sections. The effects of material strength and magnitude of axial loads were investigated. It is determined that the strength of core concrete has no effect on the response of concrete jacketed RC columns under lower axial loads while it increases the strength and reduces the ductility under higher axial load levels
Load-rating factors are used to evaluate the service life or safety of culverts based on the culvert wall strength and soil cover depth over the culvert. Current culvert load-rating methods have deficiencies in identifying potential critical conditions of corrugated metal culverts. Current load-rating procedures do not provide explicit guidance to the engineer for load rating and evaluation of culvert condition. To ensure good performance over the design life, corrugated metal culverts must be designed and regularly evaluated using an effective load-rating method. The main objective of this study was to investigate the effectiveness of current load-rating procedures. Recommendations are made to improve the analysis and evaluation procedures for corrugated metal culverts. The proposed load-rating procedure is based on an extensive review of load-rating procedures and design practices, experimental data, and theoretical investigations. The proposed method does not include a rating factor for cover depth. However, the design cover depth is required to be checked during the initial design stage to ensure structural stability. New capacity reduction factors are introduced for culvert wall and seam, which require different appraisals for wall and seam during annual inspections. The effect of external live loads is not included in the proposed load-rating procedure for deep culverts and for culverts subjected to low live load stresses. Field data from 39 in-service culverts showed that the proposed load-rating procedure is effective in evaluation of the existing condition of culverts.
This study investigates the accuracy of FEMA 356 shear and flexure modeling procedures for reinforced concrete (RC) columns and beam-column joints with poorly detailed or insufficient reinforcement. Following the FEMA 356 guidelines, generalized flexure and shear force-deformation relations were developed and compared with the experimental data from 26 column specimens and 17 beam-column joint specimens. Specifically, the measured and predicted responses were compared and evaluated: at yield displacement and the corresponding lateral load, lateral load and displacement at ultimate, and at axial load failure. In general, while the FEMA 356 models predict the lateral strength of columns reasonably well, they underestimate the shear strength of beam column joints. The predicted initial stiffness and deformations at both yield and ultimate are conservative for columns.
An unusually large number of minarets, which are slender tower structures, collapsed during the 1999 Kocaeli and Duzce, Turkey earthquakes with resulting damage to surrounding buildings and loss of life. The potential effects of the subsequently observed poor reinforcement detailing on the dynamic response is discussed. The probable cause of the extensive damage to reinforced concrete minarets is investigated by studying the observed failure modes and their seismic performance, and through the dynamic analysis of a representative minaret. The effects of spiral stairs, door openings, and balconies on the dynamic behavior are examined. The maximum dynamic internal force demands were compared with the calculated capacities. The locations of the maximum axial, shear, and flexural demands predicted from the finite element analysis of the minaret model were consistent with the earthquake damage observed at those critical locations.
